GCN LAB IMPRESSIONS

Officially extinct: NASA pulls plug on its last mainframe

Sad news today. Well, not really, but bear with me. NASA finally shut down its last mainframe computer, an IBM Z9.

The Z9 isn’t really all that old as computers go, having come out in 2005. But it was part of a dying breed. It might have been a fast and light model dinosaur, but it was pegged for extinction nonetheless.

It just makes more sense to use lots of networked, smaller computers to achieve the same results. Heck, you can get supercomputer power today by chaining a bunch of PlayStation 3’s together.

NASA CIO Linda Cureton posted a blog over the weekend talking about the really old days of computing, when mainframes used to take up an entire room. And I have to admit a little bit of nostalgia for that time period. I’m a bit too young to have worked in a place like that, but I do fondly remember a field trip in middle school to check out the computers at several federal agencies.

It was a really wild experience. I remember the rows of tapes lined up on shelves and the glass-encased computers, the size of refrigerator boxes, humming along making all types of rumbling and squeaking sounds.

I watched the technicians moving around, tending to the machines and thought they were like alchemists, the only ones who understood the formulas and complexities that kept the beast alive. And I was pretty sure that I wanted to be one of them, to join their cult of technology.

By the time college rolled around, mainframes were getting to the edge of their useful lifespan. My school had one, though it was a single box tucked away in the basement of an older, converted residence hall.

Still, I had fun learning its secrets. In just a month of poking around, I had hacked my way to admin status and was secretly running a MUD on the backend. It didn’t hurt my operation when some of the actual wardens of the machine joined the game.

But I knew mainframes were in trouble even back then. Desktop computers were hardly the powerhouses they are today, yet my souped-up 486 could almost rival the mainframe in power, while costing a heck of a lot less (I could afford it in college), taking up less space and consuming less power than the dinosaur across campus.

So by the time I started actually working, there were not many mainframes around. There was one at the newspaper I worked at after college, but I was a reporter and didn’t really get to see it too much. And I thought that hacking into the mainframe of the first company I worked for after school was a bad idea, so I pretty much avoided it altogether.

But I bet a lot of you readers did get a chance to fool around with some impressive mainframes. Tell us, what was it like? We’d all love to hear about the secret life of the mainframe, when command prompts were godlike in power, finding a bug in the machine might be meant literally, and dinosaurs still ruled the day.

If you have a good story, post it in the comments section below.
 

Reader Comments

Fri, Mar 30, 2012

I am a Systems Programmer and work on one every day, a z10, upgraded from a z9. The digging between Mainframe and LAN has turned into chickens and horses.. Many chickens to get the power of one horse. The reliability and response time can't be matched by racks of servers. Extinct? Not even close!

Fri, Feb 24, 2012 Rafael

what are they using to compute the information now??

Wed, Feb 22, 2012 Erik

My wife put me through Grad SChool as an IBM 370 operator!!

Mon, Feb 20, 2012 Miguel

I remember in college that UNH had the largest Digital VAX 870 installation in the world. It was in a room 35 x 100 and had four rows of tape drives. 15 years later that same system with 30 times the capability was the size of three refrigerators.

Fri, Feb 17, 2012

You wrote 'Heck, you can get supercomputer power today by chaining a bunch of PlayStation 3’s together.' Yes you can get the same number of instructions per second however there is a whole class of problems that you can't run on a bunch of computers and these are exactly the type of problems physics presents for example. If you have a big hairy equation that can be decomposed you can take parts of the equation and spread it around. BUT there are problems that don't decompose and these must be run on a single computer.

Show All Comments

Please post your comments here. Comments are moderated, so they may not appear immediately after submitting. We will not post comments that we consider abusive or off-topic.

Please type the letters/numbers you see above